Chris Searle wrote:
> On 7. april. 2008, at 10.10, Ralf G. R. Bergs wrote:
>
>> On 07.04.2008 10:00, Heiko Schlittermann wrote:
>>> Is there anybody else who had the same problem?
>> ACK.
>>
>> The freshclam daemon obviously locked the database long enough for
>> clamd
>> to time out retrying to obtain a lock on the DB. So it ended itself.
>
> Same here.
>
> One thing I wonder - with the following setting:
>
> av_scanner = clamd:/var/run/clamav/clamd.ctl
>
> Is there any way to configure exim to say that instead of rejecting
> when this is missing in /var/run - allow mail as normal - just add a X-
> something-sensible here header to say that mail was not virus scanned?
>
Someone asked about this kind of thing last year and I suggested they
use the clamd PING to see if it's there. They took that idea and created
an awesome little system and posted the entire thing to the mailing
list. You could use something very similar to test to see if clamd is
working and if it's not, skip or defer the malware tests.
The existence of the socket file does not mean that clamd is running,
just that it wasn't deleted if clamd crashed.
It's available in the archive here:
http://lists.exim.org/lurker/message/20070918.172526.ff9818ec.en.html
--
The Exim Manual
http://www.exim.org/docs.html
http://docs.exim.org/current/
